Loading video...

Video Failed to Load

Go Home

Fruit-fly connectome optimized inference for DeepSeek-V4.1-Flash on 2× NVIDIA GB300 DGX Stations! It impressively hits 56k tok/s prefill and 201.7 tok/s c1 decode. Well done little dude.

MrCatid's profile picture

catid

4,424 subscribers

12,293 views • 2 months ago •via X (Twitter)

Science & Technology
LIVE

Anya Rossi• Live Now

Private livecam show

0 Comments

No comments available

Comments from the original post will appear here

Related Videos

DeepSeek's open-sourced engine hits 1.7M tok/s 🐳 DeepSeek released DeepInfer, which serves V4.1 Flash in production, with 325 tok/s for one chat and about 1.7M tok/s across 64 Nvidia Blackwell GPUs Run AI models locally ->
0:12

DeepSeek's open-sourced engine hits 1.7M tok/s 🐳 DeepSeek released DeepInfer, which serves V4.1 Flash in production, with 325 tok/s for one chat and about 1.7M tok/s across 64 Nvidia Blackwell GPUs Run AI models locally ->

atomic.chat

30,817 views • 26 days ago

Okay, I guess I didn't waste all that money after all Deepseek V4.1 Flash 304 tok/s code, 135 tok/s prose Zero optimization, will have it max optimized by the morning.
0:17

Okay, I guess I didn't waste all that money after all Deepseek V4.1 Flash 304 tok/s code, 135 tok/s prose Zero optimization, will have it max optimized by the morning.

Killy

44,653 views • 2 months ago

RadixArkのQwen3.8-Flash-Next-NVFP4 つかってFreeTokenをcodexで改造してやっと良きスピード出た Activeが小さいからそれなりに速度出た RAM DDR5 8ch GPU - RTX5090 x1 Prefill 277.5 tok/s Decode 59.13 tok/s - RTX5090 x2 Prefill 619.5 tok/s Decode 83.14 tok/s
0:23

RadixArkのQwen3.8-Flash-Next-NVFP4 つかってFreeTokenをcodexで改造してやっと良きスピード出た Activeが小さいからそれなりに速度出た RAM DDR5 8ch GPU - RTX5090 x1 Prefill 277.5 tok/s Decode 59.13 tok/s - RTX5090 x2 Prefill 619.5 tok/s Decode 83.14 tok/s

ぱぷりか炒め

27,962 views • 2 months ago

DeepSeek v4.1 Flash on 3x DGX Sparks
0:18

DeepSeek v4.1 Flash on 3x DGX Sparks

Mia

31,475 views • 2 months ago

Deepseek v4.1 Flash on 24GB Macbook. I think I reached my limit, 1.36 tok/s. It is far from being usable but the feeling of having a SOTA model on my consumer machine is awesome. I started from 0.74 tok/s and almost double it. It was a great journey.
0:15

Deepseek v4.1 Flash on 24GB Macbook. I think I reached my limit, 1.36 tok/s. It is far from being usable but the feeling of having a SOTA model on my consumer machine is awesome. I started from 0.74 tok/s and almost double it. It was a great journey.

Marco Franzon

90,955 views • 2 months ago

4 DGX Sparks running DeepSeek V4.1 Flash 6 concurrent sessions, 170.73 tok/s coding throughput. 🔥 1,200 output tokens in 7.028 seconds. Warm single-stream: 💻 Code: 77.61 tok/s 🧮 Math: 74.08 🧠 Reasoning: 48.75 ✍️ Prose: 33.02 6-stream aggregate: 💻 Code: 170.73 tok/s 🔢 Counting: 291.64 tok/s 552B MoE · MXFP4 · TP4 · vLLM + DSpark + CUDA graphs ✅ Tool calling + vision ✅ 127,480-token retrieval test passed ✅ 300K configured context Video replays the actual recorded streams at 1× speed. Thinking OFF, 200-token cap per coding request. Short-prompt coding throughput Credit to Tech2Wild for the Boot9 recipe. NVIDIA AI
0:10

4 DGX Sparks running DeepSeek V4.1 Flash 6 concurrent sessions, 170.73 tok/s coding throughput. 🔥 1,200 output tokens in 7.028 seconds. Warm single-stream: 💻 Code: 77.61 tok/s 🧮 Math: 74.08 🧠 Reasoning: 48.75 ✍️ Prose: 33.02 6-stream aggregate: 💻 Code: 170.73 tok/s 🔢 Counting: 291.64 tok/s 552B MoE · MXFP4 · TP4 · vLLM + DSpark + CUDA graphs ✅ Tool calling + vision ✅ 127,480-token retrieval test passed ✅ 300K configured context Video replays the actual recorded streams at 1× speed. Thinking OFF, 200-token cap per coding request. Short-prompt coding throughput Credit to Tech2Wild for the Boot9 recipe. NVIDIA AI

Wësche

24,282 views • 2 months ago

How is DeepSeek v4.1 Flash this fast on 2 DGX Sparks??? This is a national security risk.
0:16

How is DeepSeek v4.1 Flash this fast on 2 DGX Sparks??? This is a national security risk.

Ray Fernando

280,617 views • 26 days ago

I think the CMP 170HX just became one of the weirdest local-LLM value plays. 👀 I got Qwen3.8-27B W8A16 running on a single 170HX 64GB with vLLM 0.27.1 + DFlash2: ⚡ ~120 tok/s decode 🚀 ~1.6–1.9K tok/s prefill 🧠 262K context with BF16 KV 📦 342K-token KV capacity 🎯 DFlash2 averaging 3.66 tokens/draft 💾 60.5 / 63.5 GiB VRAM used 27B model is doing ~120 tokens/sec on one GPU. Software optimization is getting ridiculous. Those numbers are from your measured W8A16 + DFlash2 run: ~120 tok/s decode, ~1,580–1,940 tok/s prefill, 342,729-token KV capacity, and 60.5/63.5 GiB VRAM usage. #170hx #LocalAI #localllm
0:11

I think the CMP 170HX just became one of the weirdest local-LLM value plays. 👀 I got Qwen3.8-27B W8A16 running on a single 170HX 64GB with vLLM 0.27.1 + DFlash2: ⚡ ~120 tok/s decode 🚀 ~1.6–1.9K tok/s prefill 🧠 262K context with BF16 KV 📦 342K-token KV capacity 🎯 DFlash2 averaging 3.66 tokens/draft 💾 60.5 / 63.5 GiB VRAM used 27B model is doing ~120 tokens/sec on one GPU. Software optimization is getting ridiculous. Those numbers are from your measured W8A16 + DFlash2 run: ~120 tok/s decode, ~1,580–1,940 tok/s prefill, 342,729-token KV capacity, and 60.5/63.5 GiB VRAM usage. #170hx #LocalAI #localllm

Own Your Compute

19,494 views • 2 months ago

nvidia/Qwen3.6-35B-A3B-NVFP4 running in vLLM nightly on my Nvidia GB10 is actually insane 50 tok/s, 4 concurrent generations. total 200 tok/s. ideal for spawning subagents or working in parallel its tool calling behavior is very good as well. I will be giving it test drive on an openclaw instance, and keep you posted More details on NVIDIA forum:
0:12

nvidia/Qwen3.6-35B-A3B-NVFP4 running in vLLM nightly on my Nvidia GB10 is actually insane 50 tok/s, 4 concurrent generations. total 200 tok/s. ideal for spawning subagents or working in parallel its tool calling behavior is very good as well. I will be giving it test drive on an openclaw instance, and keep you posted More details on NVIDIA forum:

Onur Solmaz

27,887 views • 5 months ago

DeepSeek V4 Flash Running 12 GB phone CPU only. - Model: 284B parameters - Speed: 1 tok/s - Just clever expert streaming from flash storage. -
0:12

DeepSeek V4 Flash Running 12 GB phone CPU only. - Model: 284B parameters - Speed: 1 tok/s - Just clever expert streaming from flash storage. -

Md Ismail Šojal 🕷️

28,301 views • 3 months ago

vllm-exl3 v0.3.0 is LIVE with custom native CUDA kernels for 2-bit EXL3 on NVIDIA DGX Spark GB10. GLM-5.3-Flash-EXL3-K2 jumped from 16.9 → 24.6 tok/s average single-stream decode, a +45.6% gain. Coding hit 27.6 tok/s, +85.6%. 🚀 The previous ExLlamaV3-backed path inside vLLM was leaving a lot of GB10 bandwidth on the table. So I rewrote the hot path specifically for EXL3 on Blackwell sm_121: → in-register Trellis dequantization → native fused MoE decode → power-of-two chunked prefill GEMM → parallel NVMe pre-warm Then I tested it side-by-side on physical DGX Spark hardware using my GLM-5.3-Flash-EXL3-K2 pack and live vLLM HTTP streaming. 🚀 𝗗𝗘𝗖𝗢𝗗𝗘 𝗧𝗛𝗥𝗢𝗨𝗚𝗛𝗣𝗨𝗧 Single-stream C1: Coding 14.9 → 27.6 tok/s +85.6% Prose 13.7 → 24.6 tok/s +79.3% Reasoning 18.9 → 25.1 tok/s +32.7% Summary 17.1 → 25.6 tok/s +50.0% Format 16.3 → 24.0 tok/s +47.7% Average: 16.9 → 24.6 tok/s 𝗡𝗘𝗧 𝗚𝗔𝗜𝗡: +45.6% ⏱️ 𝗙𝗜𝗥𝗦𝗧-𝗧𝗢𝗞𝗘𝗡 𝗥𝗘𝗦𝗣𝗢𝗡𝗦𝗜𝗩𝗘𝗡𝗘𝗦𝗦 Coding TTFT: 2,344 ms → 859 ms That is a 63.3% reduction, or about 2.7× faster to first token. Follow-up turn with prefix cache hit: 5,608 ms → 3,588 ms 1.56× faster. ⚡ 𝗪𝗛𝗔𝗧 𝗖𝗛𝗔𝗡𝗚𝗘𝗗 𝗢𝗡 𝗧𝗛𝗘 𝗚𝗣𝗨 40 routed-MoE layers: 19.9 ms → 11.5 ms per token Per-layer MoE compute: 497 μs → 287.8 μs That removes 8.4 ms of MoE compute from every generated token. Total per-step wall time: 59.2 ms → 40.6 ms -31.4% The key is `p2b_fused_moe`. Instead of expanding EXL3 weights through a traditional intermediate path, the new kernel performs Trellis dequantization in-register while executing the routed expert computation. The weights stay compressed until the GPU actually needs them. 🔥 𝗣𝗥𝗘𝗙𝗜𝗟𝗟 𝗚𝗢𝗧 𝗔 𝗡𝗔𝗧𝗜𝗩𝗘 𝗣𝗔𝗧𝗛 𝗧𝗢𝗢 The new `exl3_gemm` uses power-of-two chunked prefill GEMM. Measured: 7.85 TFLOPS 13.0× faster than the legacy prefill kernel 1,875 tok/s cold prefill sustained across 65K context 💾 𝗧𝗛𝗘 𝗕𝗢𝗢𝗧 𝗣𝗔𝗧𝗛 𝗡𝗘𝗘𝗗𝗘𝗗 𝗪𝗢𝗥𝗞 𝗧𝗢𝗢 Loading a ~91 GiB model is part of the user experience. Standard shard loading is mostly serial. The updated recipe parallelizes NVMe pre-warm across 8 workers so the storage controller gets used properly instead of feeding a ~100 GiB model one shard at a time. That turns boot-time storage into another optimization target instead of something we simply accept. 💡 𝗧𝗪𝗢 𝗦𝗘𝗥𝗩𝗜𝗡𝗚 𝗙𝗟𝗔𝗚𝗦 𝗪𝗢𝗥𝗧𝗛 𝗞𝗡𝗢𝗪𝗜𝗡𝗚 `--long-prefill-token-threshold 1024` Prevents giant prefill chunks from monopolizing step budgets and starving parallel decode sessions. `--enable-prefix-caching` Avoids paying for the same conversational prefix again on follow-up turns. 📦 𝗘𝗩𝗘𝗥𝗬𝗧𝗛𝗜𝗡𝗚 𝗜𝗦 𝗢𝗣𝗘𝗡 vllm-exl3: GLM-5.3-Flash one-Spark recipe: Model: This is why I like working at the kernel level. The model did not change. The quant did not change. The hardware did not change. The execution path did. 16.9 → 24.6 tok/s. 🛠️ vLLM turboderp
0:11

vllm-exl3 v0.3.0 is LIVE with custom native CUDA kernels for 2-bit EXL3 on NVIDIA DGX Spark GB10. GLM-5.3-Flash-EXL3-K2 jumped from 16.9 → 24.6 tok/s average single-stream decode, a +45.6% gain. Coding hit 27.6 tok/s, +85.6%. 🚀 The previous ExLlamaV3-backed path inside vLLM was leaving a lot of GB10 bandwidth on the table. So I rewrote the hot path specifically for EXL3 on Blackwell sm_121: → in-register Trellis dequantization → native fused MoE decode → power-of-two chunked prefill GEMM → parallel NVMe pre-warm Then I tested it side-by-side on physical DGX Spark hardware using my GLM-5.3-Flash-EXL3-K2 pack and live vLLM HTTP streaming. 🚀 𝗗𝗘𝗖𝗢𝗗𝗘 𝗧𝗛𝗥𝗢𝗨𝗚𝗛𝗣𝗨𝗧 Single-stream C1: Coding 14.9 → 27.6 tok/s +85.6% Prose 13.7 → 24.6 tok/s +79.3% Reasoning 18.9 → 25.1 tok/s +32.7% Summary 17.1 → 25.6 tok/s +50.0% Format 16.3 → 24.0 tok/s +47.7% Average: 16.9 → 24.6 tok/s 𝗡𝗘𝗧 𝗚𝗔𝗜𝗡: +45.6% ⏱️ 𝗙𝗜𝗥𝗦𝗧-𝗧𝗢𝗞𝗘𝗡 𝗥𝗘𝗦𝗣𝗢𝗡𝗦𝗜𝗩𝗘𝗡𝗘𝗦𝗦 Coding TTFT: 2,344 ms → 859 ms That is a 63.3% reduction, or about 2.7× faster to first token. Follow-up turn with prefix cache hit: 5,608 ms → 3,588 ms 1.56× faster. ⚡ 𝗪𝗛𝗔𝗧 𝗖𝗛𝗔𝗡𝗚𝗘𝗗 𝗢𝗡 𝗧𝗛𝗘 𝗚𝗣𝗨 40 routed-MoE layers: 19.9 ms → 11.5 ms per token Per-layer MoE compute: 497 μs → 287.8 μs That removes 8.4 ms of MoE compute from every generated token. Total per-step wall time: 59.2 ms → 40.6 ms -31.4% The key is `p2b_fused_moe`. Instead of expanding EXL3 weights through a traditional intermediate path, the new kernel performs Trellis dequantization in-register while executing the routed expert computation. The weights stay compressed until the GPU actually needs them. 🔥 𝗣𝗥𝗘𝗙𝗜𝗟𝗟 𝗚𝗢𝗧 𝗔 𝗡𝗔𝗧𝗜𝗩𝗘 𝗣𝗔𝗧𝗛 𝗧𝗢𝗢 The new `exl3_gemm` uses power-of-two chunked prefill GEMM. Measured: 7.85 TFLOPS 13.0× faster than the legacy prefill kernel 1,875 tok/s cold prefill sustained across 65K context 💾 𝗧𝗛𝗘 𝗕𝗢𝗢𝗧 𝗣𝗔𝗧𝗛 𝗡𝗘𝗘𝗗𝗘𝗗 𝗪𝗢𝗥𝗞 𝗧𝗢𝗢 Loading a ~91 GiB model is part of the user experience. Standard shard loading is mostly serial. The updated recipe parallelizes NVMe pre-warm across 8 workers so the storage controller gets used properly instead of feeding a ~100 GiB model one shard at a time. That turns boot-time storage into another optimization target instead of something we simply accept. 💡 𝗧𝗪𝗢 𝗦𝗘𝗥𝗩𝗜𝗡𝗚 𝗙𝗟𝗔𝗚𝗦 𝗪𝗢𝗥𝗧𝗛 𝗞𝗡𝗢𝗪𝗜𝗡𝗚 `--long-prefill-token-threshold 1024` Prevents giant prefill chunks from monopolizing step budgets and starving parallel decode sessions. `--enable-prefix-caching` Avoids paying for the same conversational prefix again on follow-up turns. 📦 𝗘𝗩𝗘𝗥𝗬𝗧𝗛𝗜𝗡𝗚 𝗜𝗦 𝗢𝗣𝗘𝗡 vllm-exl3: GLM-5.3-Flash one-Spark recipe: Model: This is why I like working at the kernel level. The model did not change. The quant did not change. The hardware did not change. The execution path did. 16.9 → 24.6 tok/s. 🛠️ vLLM turboderp

Cruz

21,784 views • 2 months ago

DeepSeek V4.1 Flash is 4.4× faster than DeepSeek V4.1 Flash
0:10

DeepSeek V4.1 Flash is 4.4× faster than DeepSeek V4.1 Flash

Adam Holter

166,519 views • 1 month ago

NVIDIA just dropped Nemotron-3-Nano:4b — a tiny 2.8GB model. Guess whose hardware runs it the fastest? - RTX 4090: 226 tok/s - RTX 3090: 187 tok/s - Mac Studio M2 Ultra: 86 tok/s - Mac Mini M4: 25 tok/s Home court advantage is real. Also trying a new layout with live performance charts. Lmk what you think!
0:13

NVIDIA just dropped Nemotron-3-Nano:4b — a tiny 2.8GB model. Guess whose hardware runs it the fastest? - RTX 4090: 226 tok/s - RTX 3090: 187 tok/s - Mac Studio M2 Ultra: 86 tok/s - Mac Mini M4: 25 tok/s Home court advantage is real. Also trying a new layout with live performance charts. Lmk what you think!

stevibe

127,672 views • 8 months ago

Day-0 Qwen3.8-27B vs Qwen3.6-27B: the voxel pagoda test. Same prompt, one attempt each, single DGX Spark, both NVFP4. 3.8: 75,291 tok · 52m 56s · 23.7 tok/s (vLLM, native MTP) 3.6: 12,117 tok · 5m 13s · 38.8 tok/s (llama.cpp, DFlash K10) 3.8 thought for 160K characters before writing a line of code.
0:15

Day-0 Qwen3.8-27B vs Qwen3.6-27B: the voxel pagoda test. Same prompt, one attempt each, single DGX Spark, both NVFP4. 3.8: 75,291 tok · 52m 56s · 23.7 tok/s (vLLM, native MTP) 3.6: 12,117 tok · 5m 13s · 38.8 tok/s (llama.cpp, DFlash K10) 3.8 thought for 160K characters before writing a line of code.

Wësche

102,744 views • 3 months ago

1,000 tok/s vs 85 tok/s visualized
0:11

1,000 tok/s vs 85 tok/s visualized

nader dabit

212,315 views • 4 months ago

Qwen3.8-Flash-Next is starting to feel like the local model Opus fans have been waiting for. Someone ran the NVFP4 176B-class Flash-Next on 2× DGX Sparks, and the results are wild. Real measured scaling → C1: 44.2 tok/s → C2: 64.6 tok/s → C4: 86.8 tok/s aggregate The per-stream speed drops with concurrency, but total throughput keeps climbing. Long-context behavior was even more impressive: → 5K: needle retrieved → 21K: needle retrieved → 84K: needle retrieved → 167K: needle retrieved → 262K: prefill succeeded, but the window was saturated That 167K retrieval test is the one I care about. Long agent runs are where models usually start losing the plot. Flash-Next didn’t. It also held up surprisingly well on physics-heavy reasoning, artifact generation, research workflows, evidence checking, and long-horizon planning. The personality is interesting too. DeepSeek V4 Flash feels like the dependable workhorse. GLM-5.2 feels like the problem-solving machine. Qwen3.8-Flash-Next feels more insightful. It has that rare ability to understand what you’re actually asking rather than just following the surface pattern. The main weakness I’ve noticed is instruction following. It can occasionally drift between prose turns where DeepSeek and GLM stay tighter. And this is why the 256GB M5 Ultra conversation gets interesting. If Apple can pair that huge unified-memory pool with enough bandwidth, this model class becomes genuinely practical for long-running local agents. We’re talking about frontier-class reasoning on hardware sitting on a desk.
0:18

Qwen3.8-Flash-Next is starting to feel like the local model Opus fans have been waiting for. Someone ran the NVFP4 176B-class Flash-Next on 2× DGX Sparks, and the results are wild. Real measured scaling → C1: 44.2 tok/s → C2: 64.6 tok/s → C4: 86.8 tok/s aggregate The per-stream speed drops with concurrency, but total throughput keeps climbing. Long-context behavior was even more impressive: → 5K: needle retrieved → 21K: needle retrieved → 84K: needle retrieved → 167K: needle retrieved → 262K: prefill succeeded, but the window was saturated That 167K retrieval test is the one I care about. Long agent runs are where models usually start losing the plot. Flash-Next didn’t. It also held up surprisingly well on physics-heavy reasoning, artifact generation, research workflows, evidence checking, and long-horizon planning. The personality is interesting too. DeepSeek V4 Flash feels like the dependable workhorse. GLM-5.2 feels like the problem-solving machine. Qwen3.8-Flash-Next feels more insightful. It has that rare ability to understand what you’re actually asking rather than just following the surface pattern. The main weakness I’ve noticed is instruction following. It can occasionally drift between prose turns where DeepSeek and GLM stay tighter. And this is why the 256GB M5 Ultra conversation gets interesting. If Apple can pair that huge unified-memory pool with enough bandwidth, this model class becomes genuinely practical for long-running local agents. We’re talking about frontier-class reasoning on hardware sitting on a desk.

FHILY👑

20,253 views • 2 months ago

This is awesome. Thanks Mia for this Was able to run with the RTX 3090 use this .env config DRAFT=mtp CONTEXT_SIZE=262144 CACHE_QUANT=8,4 GPU_MEM_GB=22 262K context: - Run 1: 65.98 tok/s - Run 2: 63.23 tok/s - Run 3: 64.41 tok/s Average: ~64.5 tok/s.
0:12

This is awesome. Thanks Mia for this Was able to run with the RTX 3090 use this .env config DRAFT=mtp CONTEXT_SIZE=262144 CACHE_QUANT=8,4 GPU_MEM_GB=22 262K context: - Run 1: 65.98 tok/s - Run 2: 63.23 tok/s - Run 3: 64.41 tok/s Average: ~64.5 tok/s.

Melvin Vivas

99,403 views • 2 months ago

🔥 Use DeepSeek-v4.1-flash for free in Orca Here's how: 1. Open OpenCode via terminal or the new Chat UI 2. Select "Deepseek-v4.1-flash-free" from the model picker 💡 Pro tip: "Deepseek-v4.1-flash-free" is 200k context only. Use it along side Codex / Claude Code to access more heavy-duty models
0:10

🔥 Use DeepSeek-v4.1-flash for free in Orca Here's how: 1. Open OpenCode via terminal or the new Chat UI 2. Select "Deepseek-v4.1-flash-free" from the model picker 💡 Pro tip: "Deepseek-v4.1-flash-free" is 200k context only. Use it along side Codex / Claude Code to access more heavy-duty models

Orca

37,944 views • 1 month ago

FREE DeepSeek V4.1 Flash UNCENSORED😳 ngl was expecting it to be gone but nodecode cyber v1 free key is still up what you get: - refused 1 of 100 JailbreakBench prompts, original refused 92 (model card) - 1M context, 150+ tok/s note: the key unlocks with a $2 top up (bot filter) and runs in their cli, the $2 stays in your balance use code 2XCYBER on that top up for 2x the tokens on the paid api try it here:
0:15

FREE DeepSeek V4.1 Flash UNCENSORED😳 ngl was expecting it to be gone but nodecode cyber v1 free key is still up what you get: - refused 1 of 100 JailbreakBench prompts, original refused 92 (model card) - 1M context, 150+ tok/s note: the key unlocks with a $2 top up (bot filter) and runs in their cli, the $2 stays in your balance use code 2XCYBER on that top up for 2x the tokens on the paid api try it here:

painn

17,693 views • 5 days ago

Two models. One DGX Spark One identical voxel Eiffel Tower prompt. Qwen3.8-27B xhigh: 83K tokens, 41m29s, 33.4 tok/s Ornith 1.5 35B-A3B: 47K tokens, 10m35s, 74.0 tok/s Thinking on, no token ceiling, matched camera. Ornith finished much faster; Qwen spent far more tokens.
0:18

Two models. One DGX Spark One identical voxel Eiffel Tower prompt. Qwen3.8-27B xhigh: 83K tokens, 41m29s, 33.4 tok/s Ornith 1.5 35B-A3B: 47K tokens, 10m35s, 74.0 tok/s Thinking on, no token ceiling, matched camera. Ornith finished much faster; Qwen spent far more tokens.

Wësche

51,527 views • 2 months ago